Decision analysis is a technique for comparing alternative courses of action when the results of the actions are not known, but the probabilities of the results can be estimated. Because the framework of decision-making under uncertainty describes much of medical practice, decision analysis is being applied to a wide variety of medical problems. Since the problems addressed are often complex, requiring the consideration of linked sequences of actions and results leading to further actions and results, the computations involved may be lengthy and difficult to carry out by hand. Preliminary experience has convinced us that a computerized aid to decision analytical calculation can be of great value in carrying out studies of the cost-effectiveness of medical procedures. We propose to develop a computer system to aid in the carrying out of clinical decision analyses. The system will consist of a set of routines written in the MUMPS programming language, which will allow a user to create a decision tree, specify and edit the probabilities and outcome values in parametric form, display and "fold back" or evaluate the tree, and perform sensitivity analysis. The system will serve as a tool for research into cost-effective medical strategies, and as a tool for teaching the techniques of cost-effectiveness analysis to medical students and physicians. In order to make the program readily usable by these potential users, we will develop documentation to allow use by those lacking previous experience with computers. We will also make the programs and documentation available to research workers who are performing medical cost-effectiveness studies at other settings.